open3d.pipelines.registration.RANSACConvergenceCriteria
- class open3d.pipelines.registration.RANSACConvergenceCriteria
Class that defines the convergence criteria of RANSAC. RANSAC algorithm stops if the iteration number hits
max_iteration
, or the fitness measured during validation suggests that the algorithm can be terminated early with someconfidence
. Early termination takes place when the number of iterations reachesk = log(1 - confidence)/log(1 - fitness^{ransac_n})
, whereransac_n
is the number of points used during a ransac iteration. Note that the validation is the most computational expensive operator in an iteration. Most iterations do not do full validation. It is crucial to controlconfidence
so that the computation time is acceptable.- __init__(*args, **kwargs)
Overloaded function.
__init__(self: open3d.cpu.pybind.pipelines.registration.RANSACConvergenceCriteria, arg0: open3d.cpu.pybind.pipelines.registration.RANSACConvergenceCriteria) -> None
Copy constructor
__init__(self: open3d.cpu.pybind.pipelines.registration.RANSACConvergenceCriteria, max_iteration: int = 100000, confidence: float = 0.999) -> None
- property confidence
Desired probability of success. Used for estimating early termination.
- property max_iteration
Maximum iteration before iteration stops.